Volatiles around two dimensional defect structures as indicated by micro diamonds in decompression cracks

Volatiles Around Two Dimensional

Defect Structures As Indicated By

Micro Diamonds In Decompression

Cracks

by

Jessika Potgieter

Dissertation submitted in fulfilment of the requirements for the degree

of

MAGISTER SCIENTIAE

in the Faculty of Natural and Agricultural Science

Department of Geology

University of the Free State

Bloemfontein

South Africa

May 2011

II

Declaration

I declare that the dissertation hereby handed in for the qualification Magister Scientiae at the University of the Free State is my own independent work and that I have not previously submitted the same work for a qualification at another University or faculty.

Signed at Bloemfontein on the ___ day of ______________ 2011.

_________________________ Jessika Potgieter

III

Abstract

New evidence exist that micro diamonds do not only form at high pressure and under high temperature conditions in the diamond window, but can also be synthesized by polycondensation of light carbon-bearing molecules at medium to low pressure conditions under favourable thermodynamic, stoichiometric and kinetic circumstances. This process may even occur close to the Earth´s surface. The studied eclogites contain OH, CO2, CO,

CH4, CH2O and CH3OH. These C:O:H-bearing volatiles can be found around totally

embedded micro cracks in nominally anhydrous minerals (NAMS). Micro cracks act like monomineralic and interphase grain boundaries, and can also be interpreted as two dimensional defect structures.

High-resolution synchrotron based FT-IR was used in the study to detect C:O:H-bearing volatiles around two-dimensional defect structures in NAMS; for example garnet. A correlation between the different C:O:H-bearing volatiles is visible in the micro diamond bearing defect structures, whereas in inclusion free defect structures, no correlation of the different C:O:H containing volatiles can be recognized. The findings from the study shows that the C:O:H-bearing volatiles, and their distribution pattern around the studied micro cracks, are indicators for the formation of micro diamonds in natural eclogites. The outcomes confirm the results from experimental studies on the growth and synthesis of diamond crystals as a consequence of polycondensation of light carbon molecules.

IV

Content

Chapter 1: Diamonds, Kimberlites and Xenoliths

1

1.1 Geotectonic Setting of Diamond Deposits 1

1.2 Kimberlites and Lamproites 3

1.3 Kimberlite Uplift 5

1.4 The Origin of Carbon 7

Chapter 2: Introduction

8

2.1 Purpose of Study 8

2.2 The Study Area 9

2.3 Previous Studies of Micro Diamonds 12

Chapter 3: Petrography

15

3.1 Sampling 15

3.2 Indicator Minerals 17

3.3 Macroscopic Rock Description 18

3.4 Microscopic Description 19

3.4.1 Minerals 19

3.4.1.1 Garnet (Fe, Mg)3Al2Si3O12 19

3.4.1.2 Omphacite (Na, Ca)(Mg, Fe, Al) Si2O6 21

3.4.2 Accessory Minerals 23

3.4.2.1 Serpentine (Mg, Fe)3Si2O5(OH)4 23

3.4.2.2 Biotite K(Mg, Fe)3AlSi3O10(F, OH)2 23

3.4.2.3 Amphibole 23

3.4.2.4 Diamond (C) 24

Chapter 4: Micro Cracks

32

4.1 Definition of Defect Structures 32

4.2 Micro Crack Description 33

4.3 Formation of Micro Cracks 34

Chapter 5: Mineral Chemistry

35

5.1 Analytical Methods 35 5.2 Minerals 36 5.2.1 Garnet 36 5.2.2 Omphacite 38 5.3 Accessory Minerals 39 5.3.1 Biotite 5.4 Conclusion 39 40

Chapter 6: Micro Diamonds

41

6.1 Introduction 41

6.2 Analytical Methods 42

6.3 Micro Diamond Types 43

6.3.1 Micro Diamonds within Garnet and Omphacite Crystals 44 6.3.2 Micro Diamonds in Micro Cracks within Garnet Crystals 46

V

6.4 Conclusion 48

Chapter 7: Volatiles around Totally Embedded Micro Cracks

49

7.1 Introduction 49

7.2 Analytical Methods 51

7.3 OH and CO2 Distribution Along Totally Embedded Micro Cracks 52

7.3.1 Correlation of OH and CO2 Distribution Profiles 52

7.3.2 Non-correlation of OH and CO2 Distribution Profiles 55

7.4 Diffusion Profiles 7.5 Conclusion

58 59

Chapter 8: Discussion and Conclusion

60

References

62

Acknowledgements

66

Appendix

67 Table 1 68 Table 2 69 Table 3 70 Table 4 71 Table 5 72 Table 6 73 Table 7 74

VI

List of Figures

Figure 1 Contrasting models showing barren and diamond-rich kimberlites. 1

Figure 2 Kimberlite pipe deposit model. 2

Figure 3 Map showing the location of kimberlite-hosted diamond mines in southern Africa.

9 Figure 4 A schematic map of a subduction model which shows kimberlite

locations hosting different xenoliths (K, K1, K2, K3 and K4) as well as lamproites (L).

10

Figure 5 Histogram comparing the isotopic composition of the diamonds from the Sloan group (Sloan Ranch, Colorado) with those from the Roberts Victor group.

11

Figure 6 Disc shaped eclogite from the Roberts Victor mine, South Africa. 15 Figure 7 Eclogite slice showing the areas selected for thin sections. 15 Figure 8 Thin sections made from the eclogite sample. a: garnet crystal, b:

omphacite crystal, c: serpentine vein, d: carbonate crystals within the serpentine vein and e: carbonate-bearing vein.

16

Figure 9 Eclogite from the Roberts Victor mine showing metasomatised rims. 18 Figure 10 Inhomogeneous garnet crystal under crossed nicols using transmitted

light surrounded by a serpentine vein.

19 Figure 11 Garnet crystals surrounded by late forming serpentine veins. As seen

under crossed nicols in transmitted light.

20

Figure 12 Omphacite crystal under crossed nicols in transmitted light showing 90° cleavage. Irregular cracks can also be seen in this figure.

21 Figure 13 Omphacite crystals surrounded by late formed serpentine veins under

transmitted light.

22

Figure 14 Serpentine vein with calcite inclusion as seen in transmitted light. 23 Figure 15 Micro diamonds in a serpentine vein as seen under reflected light. 24 Figure 16 Micro diamonds in a serpentine vein. Reflected light was used. 25 Figure 17 A cubic micro diamond in a serpentine vein in a garnet crystal. Calcite

is also observed inside the serpentine vein. Reflected light was used.

25

Figure 18 A serpentine vein inside a garnet crystal. Micro diamonds and calcite crystals can be observed inside the vein. Reflected light was used.

26 Figure 19 Micro diamond in a serpentine vein. Calcite is also observed. As seen

under reflected light.

26

VII Figure 21 Micro diamonds in a crack in garnet. These diamonds are 1 μm in

diameter. Reflected light was used.

27

Figure 22 Back scatter electron (BSE) image of a micro diamond in a crack inside a garnet crystal.

28 Figure 23 A topographical BSE image of a micro diamond in a crack within a

garnet crystal.

28

Figure 24 BSE image of micro diamonds in cracks located in a garnet crystal. 29 Figure 25 Micro diamond in a crack inside a garnet. SEM BSE image. 29 Figure 26 BSE image of a micro diamond in a crack hosted in garnet. 30 Figure 27 BSE image showing micro diamonds in a crack in garnet. 30 Figure 28 The BSE image shows two micro diamonds found in a serpentine vein. 31 Figure 29 A micro diamond in a crack within a garnet crystal. The shape of the

diamond appears to be controlled by the margins of the crack.

31

Figure 30 Defect structures in crystals. 32

Figure 31 Garnet crystal showing the XMg and XCa ratios. 37

Figure 32 Plots for pyroxenes and garnets. 38

Figure 33 A diamond peak can be observed at 1331.94 cm-1. The peaks ranging between 353.45 and 388.52 are indicative for the garnet matrix.

44 Figure 34 Micro diamond peaks can be seen at 1331.98 cm-1 and 1322.67 cm-1.

The peaks formed at 354.26 cm-1 and 389.53 cm-1 are due to the presence of the garnet host.

44

Figure 35 Micro diamond peaks are observed at 1324.27 cm-1 (red) and 1323.95 cm-1 (blue).

45 Figure 36 Garnet peaks were detected at 354.60 cm-1 and 390.04 cm-1, whereas a

micro diamond peak was observed at 1333.69 cm-1_.

45

Figure 37 One micro diamond peak is located at 1336.27 cm-1_{. 45}

Figure 38 Peaks were detected at 1331.93 cm-1 and 1322.50 cm-1. These are indicative for micro diamonds.

46

Figure 39 Only one micro diamond peak is observed at 1333.60 cm-1. 46 Figure 40 Micro diamond peaks were detected at 1331.95 cm-1 _{and 1322.59 cm}-1

respectively.

47

Figure 41 Two micro diamond peaks are also seen at 1332.01 cm-1 and 1322.67 cm-1.

VIII Figure 42 Two dimensional maps showing the distribution of OH, CO2 and other

volatiles in the studied garnets.

53

Figure 43 A: CO2 peak at 2300 – 2400 cm-1, B: OH peak at 3580 – 3740 cm-1 and

C: C:O:H bearing volatile peak at 1360 – 1560 cm-1. Examples of C:O:H bearing volatiles are CO, CH4, CH2O and CH3OH.

54

Figure 44 Two dimensional maps showing the distribution of OH and CO2 in the

studied garnets.

55

Figure 45 OH profile measured over a distance of 220 μm X 6 μm. The figure shows that OH is located in two-dimensional defect structures which act as mono mineralic and interface structures and does not occur in the crystal lattice.

56

Chap

1.1 G

Figure 1 and Gur Diamon those of are thus whereas continen Diamon mineral bearing not diam the garn Figure restricte adjacen derived upwelli lithosph root (H craton. In the B lithosph

pter 1: D

eotecton

1. Contrastin rney 1986). ndiferous ki f an Archae s formed o s lamproite ntal crust (4 nds are for ls on the K g kimberlite mond bearin net peridotit 1 shows ed to within nt mobile be d from sim ing materia here-astheno Haggerty, 19 Only kimbe Boyd and G here-astheno

Diamond

nic Settin

ng models sh imberlites a ean age. Kim

ver the ent es occur in 40-55 km) a rmed in re aapvaal cra s located w ng, which d te stability f contrasting n the bound elts. (A) In t milar depths al. Asthenos osphere bou 986). Lithos erlites whic Gurney (198 osphere bou

ds, Kimb

g of Diam

howing barre are most co mberlites ca tire time sc n cratonize and lithosph latively co aton in Sout within the Ar demonstrate field where g models i ds of the Ka the Haggert within th spheric dia undary (LA spheric diam ch pass throu 86) model, undary, the

berlites

mond De

en and diam ommonly fo an range in cale of the ed accreted here (150-25 ol continen th Africa cl rchaean cra s that below diamonds c illustrating aapvaal crato ty (1986) an e asthenosp monds are AB) in the v monds occu ugh this reg kimberlites e location o

and Xen

eposits

mond-rich kim ound in Prec age from l earth’s hist d mobile b 50 km) (Mit ntal roots. learly show aton. Kimbe w the Kaapv can be form why diam on and barr nd Mitchell phere as a formed by vicinity of t ur only with gion can ent s are derive of which is

noliths

mberlites. (M cambrian cr ate Archaea tory. Kimbe belts in reg tchell, 1986 The distrib ws the occur erlites outsid vaal craton t med. mond-bearin ren kimberl (1995) mo a result of y methane the deepest hin the harz

train xenocr ed from diff s defined b Mitchell, 199 ratons espec an to cainoz erlites are c gions of th 6, 1995). bution of i rrence of di de of the cr the lithosph ng kimberl ites are con del, kimber partial me dissociation parts of th zburgitic roo rystal diamo ferent depth by the equil 1 95, Boyd cially in zoic and cratonic, hickened ndicator iamond- raton are here is in ites are nfined to rlites are elting of n at the e craton ot of the onds. (B) hs at the libration

2 parameters of garnet lherzolite xenoliths found in kimberlites. In the mobile belts the boundary is considered to lie within the graphite stability field. Diamonds are believed to be stable only within the deepest part of the craton root. In this model all diamonds are of lithospheric origin.

1.2 K

Figure 2 Two typ Type 1 Mg-rich bulk ea (Sheaha Type 2 South A isotope Cherry,

Kimberlite

2. Kimberlite pes of kimb kimberlite h ilmenite a arth, which an and Cher kimberlite Africa. The ratios pinp , 1993).

es and La

e pipe depos berlites can b is an olivin and Cr-poor is suggesti rry, 1993). is a mica-k ey contain point the fo

amproite

it model (Sh be found: ne-monticel r Ti-pyrope ive of the f kimberlite, o Al-rich di ormation of

es

heahan and C llite-serpent e. The Sm/N formation o or so-called iopside, mi f type 2 kim Cherry., 1993 tine-kimberl Nd isotope r of type 1 ki d “orangeite icas and B mberlites in ) lite. It show ratios are cl imberlite in e” which ca Ba-K-V-Tita the lithosp ws macrocry lose to thos n the asthen an only be f anites. The phere (Shea 3 ystals of se of the nosphere found in Sm/Nd ahan and

4 Type 1 kimberlites are more CO2-rich, whereas type 2 kimberlites are more H2O-rich and

therefore type 2 kimberlites shows a higher fO2 (Sparks et al., 2006). In general kimberlites

are highly explosive extrusive rocks. These rocks travel through the lithosphere within a view hours (Sparks et al., 2006). They collect debris such as mantle xenoliths, crustal xenoliths and diamonds on their way to the surface.

The typical type 1 kimberlite can show three facies: the crater-facies, the diatreme-facies and the hypabyssic-facies (Mitchell, 1995; Sparks et al., 2006). All three complexes are about 1 km in depth. The crater-facies is mostly eroded and the formation of the facies is not yet understood. The diatreme-facies is characterized by high gas concentrations and the hypabyssic-facies shows typical magmatic fractionation processes.

Type 2 kimberlites shows nearly the same facies, but instead of the elliptical type 1 diatremes, they form thin (2m) and long (10km) fissures. Type 2 kimberlites usually form sub-parallel fissures, but only a few of them are diamond-bearing (Sheahan and Cherry, 1993).

A kimberlitic melt can be formed at low melt gradients in the upper mantle from a H2O and

CO2 (or other C-O-H species) bearing peridotites (Mitchell, 1995). These peridotites occur as

so-called phlogopite-dolomite-peridotites, phlogopite-peridotite or as dolomite-peridotite (Sommer, 2009a).

5

1.3 Kimberlite Uplift

As a kimberlite travels through the mantle the diamonds are incorporated into the kimberlitic melt. These diamonds are then carried to the surface of the earth. The survival of these diamonds depends on the fO2 of the kimberlitic melt. If the fO2 is too high during uplift the

diamonds will be resorbed due to a reaction to CO2. The diamonds will turn back in to

graphite (Fedortchouk et al., 2005 and Fedortchouk and Canil, 2009).

The addition of small amounts of H2O and CO2 will cause the formation of partial melts that

are rich in these compounds. This is due to a decrease in melt temperature of about 200°C (Sparks et al., 2006).These partial melts will rise adiabatically triggering the production of volatiles by the carbonate bearing kimberlitic melt. These volatiles are produced at depths of about 100 km. The volatile increase within the kimberlitic melt leads to the explosive stage of the kimberlite, which ends in an over-pressured choked flow (Sparks et al., 2006). These authors developed a 4-stage model for the emplacement of the kimberlite.

According to this model, kimberlite eruption began near the surface, originally from a fissure as a consequence of the magma being over-pressured due to its high volatile content. The original eruption produced a crater, but due to sustained over-pressure of the magma most of the erupted material was expelled from the crater (Sparks et al., 2006).

The second stage was the formation of the pipe caused by the widening and deepening of the crater, this stage was thus seen as an erosive phase. Stage 3 began when the crater widened to a critical point when the exploding mixture reached 1 atm. Past this point material was no longer ejected from the crater thus causing deposition within the pipe. If favourable conditions prevailed, fluidisation of the deposited pyroclastic materials could occur. The fluidisation was seen as a procedure that changed unconsolidated pyroclastic debris in the vent, and it was not suggested that this was the process that formed the pipe. The last stage involved post-emplacement hydrothermal metamorphism and alteration (Sparks et al., 2006). These stages, especially stages 2 and 3, were not imagined to be simple two-stage processes, but processes that could be repeated regularly causing overlapping events of pipe widening, emptying and filling. It is therefore implied that the formation of a pipe could be a long-lived process (Sparks et al., 2006).

An alternative model is that kimberlites can move up to Rayleigh-wave speed through the continental crust. The mechanism proposed for this alternative model is “Weertman cracks”.

6 These cracks are two- dimensional liquid-filled fluid fractures, which may potentially be used for the migration of the kimberlitic melt from the Earth’s mantle to the surface (Weertman, 1971a/b; Takada, 1990; Spence and Turcotte, 1990; Roper and Lister, 2007; Sommer et al., 2009b).

The first studies on fluid driven fractures were experiments using different liquids, water-filled crevasses in glaciers and vertical magma transport beneath ocean ridges (Weertman, 1971a/b). Recently, Sommer et al., 2009b gave new evidence which supports the formation of “Weertman” cracks as a possible mechanism for the supersonic ascent of kimberlitic melts into the continental lithosphere.

Field observations of peridotitic and eclogitic mantle xenoliths within kimberlites supports the theory for the formation of “Weertman” cracks. This is because the mantle xenoliths are incorporated into the kimberlite. As the kimberlite nears the earth’s surface these xenoliths of up to 200kg gets stuck inside the kimberlite. This directly contradicts the previously mentioned four-stage eruption model (Sparks et al., 2006). The reason for this is that these xenoliths are found within the upper part of the kimberlite and according to the Sparks model should have been expelled in the eruption. The xenoliths were thus not expelled due to eruption processes.An extremely fast crystallization process of the kimberlitic melt is also necessary to keep the heavy xenoliths from sinking deeper within the pipe.

7

1.4 The Origin of Carbon

Two models exist in explaining the origin of the carbon:

1) All carbon is primordial carbon, which was concentrated in the mantle during the accretion of the earth (Melissa et al., 1995).

2) The carbon is partially from the mantle and partially re-subducted crustal carbon (organic carbon) (Melissa et al., 1995).

Carbon isotopes can be used to locate the origin of the carbon. Carbon has two stable isotopes,

12_{C and} 13_{C. P-type diamonds show relative homogeneous C isotope values of about –5 ‰.}

E-type diamonds, in contrast show a significant variation in C-isotopes, indicating an inhomogeneous mantle source or a combination of sources of carbon, e.g. subduction. In general E-type diamonds are younger than P-type diamonds (Melissa et al., 1995). This is due to plate tectonic processes, which brings about the distribution of biogenic carbon in the sublithospheric mantle.

8

Chapter 2: Introduction

2.1 Purpose of Study

Diamonds are thought to be formed at high pressures and temperatures in the diamond window. Two types of diamonds can occur: E- and P-type. E type diamonds occur with eclogites and are formed from organic carbon, whereas P type diamonds are associated with peridotites and are formed from inorganic carbon. Resorbtion experiments on diamonds gave evidence that diamonds start to resorb within a few minutes to a couple of hours depending on the fO2. The conclusion formed from these experiments is that uplift from the diamond-

bearing host must be within 10 minutes to about 5 hours if the diamonds are formed in the Earth´s upper mantle (Fedortchouk et al., 2005 and Fedortchouk and Canil, 2009; Arima and Kozai., 2008). Rudenko et al., 1993 identified these micro diamonds within micro cracks and in late-formed serpentine veins and found in his experiments that micro diamonds can be formed at very low pressures.

The purpose of this study therefore is to find evidence for the formation of micro diamonds close to the Earth’s surface. Volatile distribution around diamond-bearing and non-diamond- bearing micro cracks was studied. Furthermore micro Raman studies was used to prove that the inclusions found in the micro cracks and the serpentine veins are micro diamonds. Eclogites from the Roberts Victor mine were used for this study.

2.2 Th

Figure superimp Zimbab Archaea Namaqu Proteroz units are IIc: Piet Southern Victor m

he Study

3. Map sho posed on the bwe craton; an Angolan ua–Natal belt zoic crust — e: Ia: Tokwe tersburg terra n marginal z mine (K) is sh

y Area

owing the l e structural u II: Archaea craton; V: t; VII: Early — Damara pr e terrain; Ib: ain; IId: We zone; Va: Kh hown as a gr location of units of Griffi an Kaapvaal Early Prot y-middle Pro rovince; IX: North-weste estern terrain heis fold be roup 2 kimbe kimberlite-h fin et al. (200 l craton; III terozoic cru oterozoic cru Saldanian p ern terrain; II n; IIIa: Centr lt; Vb: Okw erlite (Field e hosted diam 03b). The stru I: Archaean ust; VI: Ear

ust — Rehob province. Th Ia: South-Ea ral zone; III wa inlier; Vc et al., 2008) mond mines uctural units Limpopo m rly-middle P bothian subp he subdivisi astern terrain Ib: Northern : Makondi f in southern are: I: Archa micro contin Proterozoic province; VI ons of the s n; IIb: Centra marginal zo foldbelt. The 9 n Africa aean nent; IV: crust — III: Late-structural al terrain; one; IIIc: e Roberts

The Ro provinc with tw Beaufor erosion found i system 3) (Skin method Ma (Sm Figure 4 xenolith lherzolit model. T oberts Victo ce of South wo small p rt Sandston levels. On in the dyke (Wagner., nner., 1989 ds resulting mith et al., 1 4. A schemat hs (K, K1, K2 te, subducted These xenoli or diamond Africa. Thi pipes (Gurn e as well as nly lithic cla es. Several 1914) and t 9). The phlo in two date 985) tic map of a , K3 and K4) d eclogite, g ths are all fo mine is sit s kimberlite ney and Ki s Karoo bas asts from th varieties o the occurren ogopites fou es, namely subduction m as well as la garnet harzbu ound in the R tuated about e deposit co irkley., 199 salt xenolith he immedia of kimberlit nce has bee

und in the 127±3 Ma

model which amproites (L urgite and lit Roberts Victo t 40 km ea onsists of tw 96). The pi hs which we ate wallrock te have bee en classified kimberlite (Allsopp an h shows kimb ). Different d thospheric e or mine (K2) st of Bosho wo dykes, on ipes contai ere derived k and deep en observed d as a Group have been nd Barrett., berlite locati diamond-bea eclogite) are (Sheahan an of in the Fr ne being as n large ma from above per lithologi d in the pi p-2 kimberl dated usin 1975) and ions hosting aring xenolith also depicte nd Cherry., 1 10 ree State sociated asses of e current ies were ipe-dyke lite (Fig. g Rb-Sr 128±15 different h (garnet ed in this 1993).

Roberts diamon lithosph contain purpose Figure 5 Ranch, C Carbon to illust subduct Diamon whereas (Meliss 13_{C valu} suggest formed mine (F s Victor is nd bearing x heric eclogi s all the a e of this stud 5. Histogram Colorado) wi isotopes ar trate the ra tion zones, nds formed s diamonds sa et al., 199 ues of 2 to tive of organ by organic Fig.5). also a man xenoliths ex ites (Fig. 4) above ment dy are of su comparing t ith those from

re used to id ange of org , while ino by organic s formed by 95) -10 are ind nic carbon c carbon. T ntle xenolith xist; garnet l ). It can be tioned diam ubducted ecl the isotopic c m the Robert dentify the o ganic carbon organic car carbon usu y inorganic dicative of (Fig.5). A 1 This is seen h-rich mine lherzolite, s seen in fig mond-bearin logitic comp composition ts Victor gro origin of dia n. Organic rbon is tho ually show a c carbon wi inorganic c 13_{C isotopic} in the eclo (Wagner, ubducted ec gure 4 that ng xenolith mposition (M of the diamo oup (Melissa amonds. Th carbon is ought to c an inhomog ill show a carbon, whe c value of -1 ogitic diam 1914). Four clogite, garn the Robert hs. The sam Melissa et al. onds from th et al., 1995) he histogram carried into ome from eneous isot homogeneo ereas value 16 is thus ty onds from r different t net harzbur ts Victor mi mples used ., 1995) he Sloan grou ) m in figure 5 o the mantl the mantl ope pattern ous isotope es of -12 to ypical for di the Robert 11 types of rgite and ine (K2) for the up (Sloan 5 is used le along e itself. n (Fig 5), e pattern -30 are iamonds s Victor

12

2.3 Previous Studies of Micro Diamonds

Micro diamonds as well as larger diamonds are known to be formed at high pressure and temperature conditions in the diamond window by polymorphic phase transformation of graphite into diamond (Fig.4). The temperature and pressure condition where this transformation takes place is at about 1000 to 1100°C and 45 to 50 kbar. Micro diamonds can also be formed due to shock metamorphism in the gas phase as seen in the Nördlinger Ries, Germany (Posges, 2005). Experimental studies on the formation of micro diamonds indicate that these types of diamonds can also be formed at medium to low pressure conditions. This process may even occur close to the Earth´s surface by the transformation of carbon-containing molecules into diamond (Rudenko et al., 1993; Rudenko and Kulakova., 1996). These authors developed a model of non-equilibrium growth of diamonds in a macroscopic open catalytic system. All known active catalysts can be found in a kimberlite. Kimberlites originate from deep fractures in the Earth´s mantle. A Kimberlite retains a large number of gas channels and these different gases are involved in allochemical reactions which take place inside the pipe. Physically spoken, the most advantageous range of micro diamond formation is defined by a certain atomic ratio of C:O:H atoms in the original gas phase or plasma and the according temperature (Bachmann et al., 1991).

From a chemical point of view, the synthesis of micro diamonds and other carbon substances have different types of carbon-carbon bonds and as a result form polycarbons. These polycarbons are products of different chemical reactions (Rudenko and Kulakova., 1993). Polycarbons like diamonds, graphite and carbon are condensation products, created in a polycondensation process. If polycondensation under favourable conditions takes place, diamonds can be formed in the gas/plasma phase at low to medium pressure conditions according to the following reactions (Rudenko and Kulakova., 1996).

2CO = CDi + O2 (a)

CO + H2 = CDi + H2O (b)

CH4 + CO2 = 2CDi + H2O (c)

CH2O = CDi + H2O (d)

CH4 = C The pol critical catalyti can be d K + nA A repr polyme number are diam remove The per of light describe system and the 1993). If equil become formati continu kimberl CDi + 2H2 lycondensat diamond n c system no described by A = {An}K  esent a ca eric condens r of molecu mond cryst ed during po rmanent add t molecules ed in equat and can thu rmodynami librium con es stagnant, on of serp uously and lite close to tion proces nucleus and ot in equilib y following  {An - mL arbon-bearin sation produ ular condens tals and L olycondensa dition of lig s L (H2O, tion ((g) an us be descri ic condition ndition in th then the fo pentine and therefore d o the Earth s is govern secondly th brium is req g equations: }K + mL ng molecul uct, which b sation, {An describes ation (Ruden ght carbon b CO2) lead nd (h)). Ki ibed by equ ns according he system (h ormation of d carbonate disequilibriu h´s surface. ned by two he formatio quired to for le, which became abs - mL}K ar light molec nko and Ku bearing mo ds to dise imberlites c uation (h) to g to the rea h) is reache f diamonds e bearing um can be Carbonate (f) critical fac on of the di rm diamond (g) is a polym sorbed by th re the solid cules such ulakova., 19 (h) olecules A a equilibrium can act as o form diam actions (a) – ed, meaning will stop im minerals th expected u and serpen ctors: First, iamond crys ds in the ga meric mono he catalytic condensatio as H2O an 996).

and the subs in the ope a macrosco monds under – (f) (Ruden g A = L or mmediately. he C:O:H until the em ntine formin the format stal itself. A s/plasma ph omer, {An} c converter, on products nd CO2, wh sequent sub en catalytic opic open c r favourable nko and Ku r the whole . Due to pe ratio will mplacemen ng reaction 13 ion of a An open hase and }K is a n is the s, which hich are btraction c system catalytic e kinetic ulakova., e system rmanent change t of the ns in the

14 kimberlitic melt therefore has a strong influence on the C:O:H ratio in the gas/plasma phase. Allochemical reactions will take place as a result of the accumulation of H2O and CO2 with

the surrounding material. Carbonate and serpentine reactions act as a catalytic converter causing the formation of micro diamonds in serpentine and carbonate bearing veins. These veins crosscut the mantle xenoliths (Rudenko et al., 1993; Rudenko and Kulakova., 1996).

Chap

3.1 Sa

The ec measure had me forming Figure 6 Figure 7

pter 3: P

ampling

clogitic sam ed 40 cm x etasomatise g-minerals i 6. Disc shape 7. Eclogite sl

Petrograp

mples collec 20 cm x 10 d rims and in eclogites. ed eclogite fr ice showing

phy

cted from 0 cm (Fig. 6 d contained .

rom the Robe

the areas sel

the Rober 6) and were d garnets a

erts Victor m

lected for thi

rts Victor d e ellipsoidal and ompha mine, South A in sections diamond m in shape. T cites that a Africa mine, South The chosen are the ma 15 h Africa samples ain rock

The sam sections mineral green o carbona Figure 8 serpentin mples were s (Fig. 8) f ls can clear mphacites a ate minerals 8. Thin secti ne vein, d: ca e then cut i from which rly be seen and dark co s. ions made f arbonate cry into slices ( h four were in the thin oloured serp

from the ecl stals within t (Fig. 7). Th e selected f sections. T pentine vein ogite sample the serpentin hese slices for detailed These miner s. The serpe e. a: garnet ne vein and e were furth analysed f rals include entine veins crystal, b: o e: carbonate-b

her cut into for this the e brownish s also conta omphacite cr bearing vein 16 28 thin sis. The garnets, ain white rystal, c: n.

17

3.2 Indicator Minerals

Indicator minerals are minerals, which due to weathering processes get washed out of the kimberlites/lamproites. These minerals such as garnet are unaffected by alteration processes and can be found in sediments. They have a distinguishing mineral chemical composition, which can be used as a diamond tracer in kimberlites/lamproites.

A Mineral that is an ideal indicator is a garnet with a high XMg ratio as well as a high Cr and a low Ca concentration (“G10-Grt”). Ca-rich garnet-lherzolites (“G9-Grt”) cannot be used as indicators for diamonds. High Mg and Cr concentrations in chromites can also be used as a diamond indicator. E-type diamonds are always associated with Na- and Ti- rich garnets (melanite component) (Gurney et al., 1993).

3.3 M

Figure 9 The stu sample garnets (green) clinopy the cent

Macroscop

9. Eclogite fr udied eclogi vary from (dark blue) in the cente yroxenes are ter of the stu

pic Rock

om the Robe ite shows a magnesium ) at the rims er to calcium e of omphac udied samp

k Descrip

erts Victor m wide varia m-rich Pyrop s (Fig 9). Th m-rich diop citic compo ple.

ption

mine showing ation in min pes (red) in he clinopyro pside (yellow osition. Recr g metasomati neral chemi n the center oxenes also w) at the rim rystallized p

ised rims (Po

stry. The g r to more h vary from ms of the sa pyrope (ligh otgieter, 2009 garnets foun hydro-grossu sodium-rich ample (Fig. ht blue) is f 18 9). nd in the ular-rich h jadeite 9). Both found in

3.4 M

3.4.1 M

3.4.1.1

The gar They ar and con Pyroxen signs of (Fig. 11 Figure 1 serpentin

Microscop

Minerals

Garnet (F

rnets (Grt) s re isotropic ntain micro ne inclusion f alteration. 1). 10. Inhomoge ne vein.

pic Descr

e, Mg)

3

Al

2 seen in the s c and have cracks (Fig ns have als . The garne eneous garne

ription

2

Si

3

O

12 studied eclo a high refra g. 10). Some so been obs ts are often et crystal und ogite sample active index e of these m served in th n surrounded der crossed n e are pale b x. The crys micro cracks he garnet c d by late fo nicols using rown to red stals are euh s may conta crystals. Th orming serp transmitted l ddish in thin hedral to su in micro dia he garnets s pentine (Ser light surroun 19 n section. ubhedral amonds. show no rp) veins nded by a

Figure 1 in transm

11. Garnet cr mitted light

rystals surrouunded by latte forming seerpentine veiins. As seen under crosse

20

3.4.1.2

The om shape o crystals omphac omphac Figure 1 cracks c

Omphacit

mphacite (O of the crysta s. Second o cites observ cite crystals 12. Omphacit an also be se

te (Na, Ca)

mph) in thi als is anhedr order interf ved in this may also b te crystal und een in this fig

)(Mg, Fe, A

is sample a ral. A distin ference col sample also be surrounde der crossed n gure

Al) Si

2

O

6 appears pale nctive 90° c lours can b o exhibited ed by serpen nicols in tran e green to c cleavage can be seen in irregular c ntine veins. nsmitted ligh colourless i n easily be thin sectio cracks. Fig. . ht showing 90 n thin secti seen in som ons (Fig. 1 13 shows 0° cleavage. 21 ion. The me of the 12). The that the Irregular

Figure 113. Omphacitte crystals suurrounded byy late formedd serpentine vveins under ttransmitted li

22

3.4.2 A

3.4.2.1

The ser brown other m Figure 1

3.4.2.2

Minor a in the pseudoh crossed

3.4.2.3

Amphib color. N

Accessory

Serpentin

rpentine (Se to black in minerals such 14. Serpentin

Biotite K(

amounts of studied th hexagonal c d polarised l

Amphibol

bole occurs No further st

y Minerals

e (Mg, Fe)

erp) crystals n thin sectio h as calcite ne vein with c

(Mg, Fe)3A

biotites wh hin section crystals. It c ight.

le

s as an acce tudies wher

)

3

Si

2

O

5

(OH

s seen in thi on (Fig. 14) (Cal). calcite inclus

AlSi3O10(

here found in ns and also can also be essory min re done on a

H)

4 s sample ar ). They are sion as seen

F, OH)2

n the studie o exhibits identified b eral in the amphibole. e massive in e also veine in transmitte ed sample. T perfect ba by the gnarl studied sam n habit. The ed. Some a ed light This minera asal cleava led bird’s ey mple. This ese mineral are intergrow al appears b age. Biotite ye extinctio mineral is 23 s appear wn with brownish e forms on under dark in

3.4.2.4

The dia diamon within t 18, 19 a individu on the s diamon The sha the wid usually electron Figure 1

Diamond

amonds foun nds occur w the garnet a and 28) as w ual garnet a setting. The nds that occ ape of the di dth and/or le well shape n pictures w 15. Micro dia

(C)

nd in the stu ithin three and omphac well as in th and omphac e diamonds cur in the se iamonds tha ength of the d and occur were taken w amonds in a s udied eclog different se cite crystals he micro cra cite crystals found in th erpentine v at were obs e crack itsel r as bright w with the SEM

serpentine ve gite sample ettings in th s (Fig. 20), acks (Figs. 2 . The size o he garnet an eins and w erved withi lf (Figs.21, white cubes M at the Seo ein as seen u are small (5 e collected in the serp 21, 22, 23, 2 of these dia nd omphaci within micro n the micro 24, 25, 26, under refle oul Nationa under reflecte 50 to 1μm) samples. D entine vein 24, 25, 26, 2 amond cryst ite crystals o cracks in o cracks mig 27 and 29) ected light. A al University ed light in diamete Diamonds ar ns (Figs. 15, 27 and 29) f tals vary de are bigger the garnet ght be contr . The diamo All the back y in South K 24 r. These re found , 16, 17, found in epending than the crystals. rolled by onds are k scatter Korea.

Figure 1 Figure 1 inside th 16. Micro dia 17. A cubic he serpentine amonds in a s micro diamo e vein. Reflec serpentine ve ond in a serp cted light wa ein. Reflecte pentine vein as used ed light was u n in a garnet used

crystal. Callcite is also

25

Figure 1 observed Figure 1 18. A serpen d inside the v 19. Micro dia ntine vein in vein. Reflect amond in a se nside a garn ted light was

erpentine vei net crystal. M used in. Calcite is Micro diamo s also observ onds and ca ed. As seen u alcite crystal under reflect 26 s can be ted light

Figure 2 Figure 2 was used 20. Micro dia 21. Micro dia d amond in a g amonds in a arnet crystal crack in garn l as seen und net. These di der reflected l iamonds are light

1 μm in diammeter. Reflec

27

Figure 2 Figure 2 22. Back scat 23. A topogra tter electron aphical BSE (BSE) image image of a m e of a micro micro diamon diamond in a nd in a crack a crack insid k within a gar de a garnet cr rnet crystal 28 rystal

Figure 2 Figure 2 24. BSE imag 25.Micro diam ge of micro d mond in a cr diamonds in rack inside a cracks locate garnet. SEM ed in a garne M BSE image et crystal e 29

Figure 2 Figure 2 26. BSE imag 27. BSE imag ge of a micro ge showing t o diamond in two micro di n a crack host amonds in a ted in garnet crack in gar t rnet 30

Figure 2 Figure 2 be contr 28. The BSE 29. A micro rolled by the image show diamond in margins of t ws two micro a crack with the crack diamonds fo hin a garnet c ound in a serp crystal. The pentine vein

shape of thee diamond ap

31

Chap

4.1 D

Micro c mineral dislocat the gar characte Figure 3

pter 4: M

efinition

cracks are t lic grain bo tion, edge d rnet crystals erized by hi 30. Defect str

Micro Cr

n of Defec

two dimens oundaries. T dislocation a s are exam igh dislocat ructures in cr

racks

ct Structu

sional defec Three types and lineage mple of line tions which rystals (Buch

ures

ct structure s of two dim structure (F eage structu facilitate vo her, 2002) s. They can mensional d Fig. 30). Th ure defects. olatile mov n also be in defect struc he studied m These line ement in th nterpreted a ctures occur micro crack eage structu he crystal ma 32 as mono r: screw ks within ures are atrix.

33

4.2 Micro Crack Description

The micro cracks in the studied eclogite are examples of lineage structures. They occur at the rims of the garnet and omphacite crystals. These micro cracks are not aligned and are found randomly in the studied minerals. In some cases they are interconnected to each other, but generally they occur isolated in the crystal matrix. These cracks are 0.1 to 0.3 μm wide and can be up to 100 μm long.

The micro cracks form between 50 and 60 kbar and at temperatures of ~ 1200°C. They act as pathways for fluid during uplift of the kimberlite and it is due to this fluid movement and chemical interaction that diamonds are formed in the cracks (Fig. 20). However not all the micro cracks contain micro diamonds, non-diamond bearing cracks are also widespread throughout the sample.

34

4.3 Formation of Micro Cracks

The formation of these micro cracks is due to decompression of the xenoliths, which takes place in the Earth´s upper mantle. Due to the fast ascent of the kimberlite through the Earth´s upper mantle and lithosphere, the density contrast between the kimberlitic melt and the xenoliths within in the kimberlitic melt changes continuously. The difference in density between the xenolith and the kimberlitic melt in the Earth's mantle is much higher than at the Earth´s surface. The density contrast at a depth at 150 to 160 km is in the order of magnitudes higher than at the Earth´s surface (Sommer et al 2009b).

The kimberlitic melt contains large quantities of volatiles. These volatiles mainly consist of H2O and CO2. Kimberlites are usually very CO2 rich. This increase in CO2 leads to an overall

viscosity of less than 1 (the same viscosity as water). In general a kimberlitic melt has a viscosity of 0.5 (Sparks et al., 2006). When the kimberlite picks up the xenolith in the Earth´s upper mantle the kimberlitic melt has the consistency of a plasma or gas. As eruption of the kimberlite takes place, water and CO2 is constantly being removed from the kimberlitic melt

to form minerals such as serpentine and calcite. Consequently the density of the kimberlitic melt will change continuously, as a result of cooling. This is the reason for the decrease in density contrast between the kimberlitic melt, the country rocks and the xenoliths within the kimberlite during the uplift. The decompression cracks are thus formed in depth, where the highest density contrast occurs (Sommer et al., 2009b).

35

Chapter 5: Mineral Chemistry

5.1 Analytical Methods

28 thin sections of the studied eclogite were investigated using transmitted light microscopy and a scanning electron microscope (SEM). Four thin sections were then selected for detailed analyses. Mineral analyses were carried out in Graz using a JEOL 6310 SEM equipped with a LINK ISIS energy-dispersive system and a MICROSPEC wavelength dispersive system. Accelerating voltage was 15 kV and 5 nA. Matrix corrections for silicates were made using the ZAF procedure, and natural mineral standards were used for calibration. Detection limits varied 0.1 to 0.5 wt. % for the JEOL SEM with the LINK ISIS energy-dispersive system.

36

5.2 Minerals

5.2.1 Garnet

Core and rim measurements of the garnet crystals were taken in order to analyse the XMg, XCa

and XFe variations present in the eclogite sample. An example of these measurements can be

seen in figure 31.

The formula used for these measurements is: XMg

ൌ

୑୥ ୑୥ା୊ୣାେୟ : XFe

ൌ

୊ୣ ୑୥ା୊ୣାେୟ : XCa

ൌ

େୟ ୑୥ା୊ୣାେୟ

The XMg varies from 0.596 at the rim to 0.550 units per formula at the centre of the sample.

These measurements thus show that the XMg is higher at the rims of the garnet sample than in

the middle.

The XFe varies from 0.322 units per formula in the centre of the garnet crystal to 0.259 units

per formula at the rim of the sample. This is in contrast to the XMg measurements due to the

FM exchange.

The XCa varies from 0.150 units per formula in the centre to 0.127 units per formula at the

rim of the studied garnet sample. These results thus indicate that the XCa is higher at the

centre of the garnet crystal than at the rims.

Figure 3 The garn 31. Garnet cr net crystal m rystal showin measures 5mm ng the XMg a m and 223 po nd XCa ratios oints were an s. Measured nalysed

with SEM wwith EDX an

37

5.2.2 O

Core an that the case in nature a visible atomic and 0.7 Figure 3 contains almandin The py pyroxen pyroxen centre o took pla

Omphacite

nd rim meas e studied om the garnet and varies b in the XAl, units per fo 5 units per 32. Plots for s diopsides ne/pyrope co yroxenes fou nes located nes located of the samp ace in the sa

e

surements o mphacites w t samples. T between 0.7 , XCa and X ormula, XNa formula (re r pyroxenes as well as omposition und in the in the centr in the rims ple have an ample: 2Ca of the omph were homog The XMg o 79 and 0.82 XNa concent a between 0 fer to appen and garnets omphacites. rims of th re of the sam s of the sam omphacitic = Al + Na. hacites wher geneous and f the studie 2 atomic un trations me .05 and 0.3 ndix). s. The pyrox . The garne he eclogite mple are mo mple are dio c compositio

re analysed d showed n ed mineral nits per form easured. XA 1 units per xene plot in ets located i sample (Fi ore Al and N opsides whi on (Fig. 32 . These mea no chemical samples ar mula. A stro Al varies bet formula and ndicated that in the eclog ig.9) are C Na-rich. Thi

ile the pyro ). The follo asurements l zonation a re homogen ong dissimi tween 0.09 d XCa betwe t the studied gite sample a-rich whe is indicated oxenes foun owing react 38 showed as is the neous in ilarity is to 0.25 een 0.58 d sample have an reas the d that the nd in the tion thus

39

5.3 Accessory Minerals

5.3.1 Biotite

The biotites located in the eclogite sample have an XMg between 0.76 and 0.79 units per

formula which is indicative for phlogopites. Refer to the appendix for the micro probe analyses.

40

5.4 Conclusion

The results of the core to rim measurements of the garnets in the eclogite sample (Fig.31) showed that the changes in the XMg and XFe ratios may be due to the sample undergoing

metasomatism and thus being heated from the rims to the centre during uplift. The change in the XCa ratios from the core to the rim of the measured garnet crystals of the eclogite sample

demonstrates that the sample was formed at high pressure conditions. Since the measurements decrease from the centre to the rim it can be deduced that the pressure dropped while the eclogitic xenolith was brought to the surface of the earth by the kimberlite (Deer et al., 1992).

It was discovered that the studied eclogite sample contained diopside as well as omphacite. The omphacite crystals were analysed and strong variations in the XCa, XNa and XAl ratios

were discovered. These variations are thought to be due to decompression (decrease in pressure) caused by the uplift of the eclogite xenolith. The diopsides discovered in the rims of the studied eclogite (Fig.9) sample are indicative of high temperature conditions and the omphacites found in the centre of the sample shows that the eclogite was formed under high pressure conditions (Deer et al., 1992).

41

Chapter 6: Micro Diamonds

6.1 Introduction

The high-pressure modification of carbon is known as a diamond. When a diamond is well crystallized, they occur as octahedrons, but cubes and cube-octahedral forms are also common, whereas rhombododecahedrons are very rare. Cubes and cube-octahedrons are formed at higher pressures and lower temperatures. Diamonds from eclogites and peridotites from the same deposit are formed under similar conditions. Typical depth ranges between 150 and 200 km and temperatures below 1200 °C. Some E-type diamonds are formed at depth of about 400 km (Spear., 1993).

The micro diamonds found in the studied eclogite sample can be divided into three groups. They are micro diamonds found in the garnet and omphacite crystals itself, micro diamonds in the micro cracks within the garnet crystals and micro diamonds in the serpentine veins. The micro diamonds found within in the garnet and omphacite crystals are usually larger than the other types of diamonds found in the sample. This is due to the fact that they were formed at greater depths and temperatures in the mantle. They thus had more time to form bigger diamonds. These micro diamonds are also well shaped. It is thought that the micro diamonds found in the micro cracks within the garnet crystals were formed at shallower depths and cooler temperatures, thus forming smaller crystals. The serpentine veins contain the most micro diamonds. These diamonds are larger than those found in the micro cracks. They are also better shaped.

42

6.2 Analytical Methods

Diamonds were identified by using a dispersive DXR Raman spectroscopy (Thermo Scientific) at the Tectonophysics Laboratory of the School of Earth and Environmental Sciences (SEES), Seoul National University.

The Raman microscope was equipped with a 532-nm laser, a standard resolution grating (5 cm-1 nominal resolution, 50-3550 cm-1 spectral range), and an optical microscope (Olympus, 50x objective). The Raman spectrum was obtained from 30 μm beneath the surface and around the embedded crack. An exposure time of 22 seconds and a laser power of 10 mW was used.

43

6.3 Micro Diamond Types

According to a study done by Rudenko et al., 1993 (see Chapter 2.3), micro diamonds can be formed at medium to low pressure conditions in serpentine veins and in two dimensional defect structures such as micro cracks. Different chemical reactions may lead to the formation of differently shaped micro diamonds. It is also thought that the chemical reactions will differ depending on the location inside the studied eclogite sample (Fig.6) for example within the garnet or omphacite crystals (Fig.20), the serpentine vein (Fig.18) or the micro crack within a garnet crystal (Fig.27). Micro Raman analysis was thus done in order to determine if shifts can be observed in the Raman spectra from these three different locations.

6.3.1 M

These m sample. garnet m diamon 1325cm some ca diamon one mic Figure 3 and 388 Figure 3 354.26 c

Micro Dia

micro diamo . The Rama matrix/cryst nds were fou m-1, while a ases both pe nd peaks we cro diamond 33. A diamon .52 cm-1 _are 34. Micro dia cm-1 _{and 389}

amonds wi

onds are fou an peaks ob tal hosting t und in the another diam eaks were o ere observed d peak can b nd peak can indicative fo amond peaks 9.53 cm-1 _are

ithin Garn

und within bserved bet the micro d samples an mond peak observed in d at 1331.98 be seen (fig be observed or the garnet s can be seen due to the pr

net and Om

garnet and tween 350 iamonds. T nalysed. On was observ the same s 8 cm-1 and 1 g. 33, 35, 36 at 1331.94 c t matrix/cryst n at 1331.98 resence of th

mphacite C

omphacite cm-1 and 3 Two peaks th ne peak was ved at abou amples as c 1322.67 cm 6 and 37). cm-1_{. The pe} tal 8 cm-1 _{and 13} he garnet hos

Crystals

crystals in t 390 cm-1 ar hat are sugg s found at a ut 1331 cm can be seen -1_{. In most c} aks ranging 322.67 cm-1_. st. the studied re indicative gestive of th about 1322 m-1 to 1336 in figure 3 cases howev between 353 The peaks f 44 eclogite e of the he micro cm-1 to cm-1. In 4 where ver only 3.45 cm-1 formed at

Figure 3 Figure 3 was obse Figure 3 35. Micro dia 36. Garnet pe erved at 133 37. One mic amond peaks eaks were de 3.69 cm-1 cro diamond s are observe etected at 354 d peak is lo d at 1324.27 4.60 cm-1 _and ocated at 133 7 cm-1 _{(red) a} d 390.04 cm -36.27 cm-1 and 1323.95 c -1_{, whereas a} cm-1 _(blue) a micro diam 45 ond peak

6.3.2 M

The mi 1331 cm in the g Figure 3 diamond Figure 3

Micro Dia

icro diamon m-1. Howev garnet crysta 38. Peaks w ds 39. Only one

amonds in

nd shifts fo ver, these pe als. They ar were detected micro diamo

Micro Cr

ound in thes eaks are not

e also narro d at 1331.93 ond peak is o

acks withi

se samples ticeably sm ower than th cm-1 _{and 13} observed at 1

in Garnet

shows pea maller that th he abovemen 322.50 cm-1 1333.60 cm-1

Crystals

aks at 1322 he micro dia ntioned pea . These are 1 2 cm-1 _and/o amond peak aks. indicative f 46 or about ks found for micro

6.3.3 M

Two mi were lo than the observe Figure 4 Figure 4

Micro Dia

icro diamon ocated at ab e micro dia ed in figures 40. Micro dia 41. Two micr

amonds in

nd Raman p bout 1322 c amond peak s 40 and 41 amond peaks ro diamond p

n Serpentin

peaks were o cm-1 and 133 ks measured were cause s were detect

peaks are als

ne Veins

observed in 31 cm-1. Th d in the gar ed by the su ted at 1331.9 o seen at 133 n the serpent hese peaks rnet crystals urrounding s 5 cm-1 _{and 1} 32.01 cm-1 _an

tine vein sam are also sm s. The other serpentine m 322.59 cm-1 nd 1322.67 c mples. Thes maller and n r peaks that minerals. respectively cm-1 47 se peaks narrower t can be y

48

6.4 Conclusion

The micro diamonds that were observed in the garnet crystals shows peaks at about 1322cm-1 and 1332 cm-1 _{. These peaks are large as well as wide. The micro diamonds are considered to}

be high pressure diamonds formed by polymorphic phase transition processes that occur in the diamond window. These high pressure micro diamonds have two crystallographic systems thus leading to the detection of these two micro diamond peaks.

Micro cracks that occur in cracks within garnet crystals were also measured. The Raman peaks caused by the micro diamonds found in these cracks are smaller and narrower than the peaks formed by the micro diamonds in the garnet crystals. The same observation was also made for the peaks formed by the micro diamonds present in the serpentine veins. These micro diamonds are always surrounded by serpentine. Since these serpentine minerals are formed close to the earth’s surface at low pressures and the peaks from the serpentine vein and the micro cracks are similar, it suggests that they were formed by the same processes at the same pressure and temperature conditions as well as at a similar time.

These results thus show that at least two generations of micro diamonds, one high pressure modification and one low pressure modification, were identified in the studied eclogite sample. The high pressure micro diamonds were formed by polymorphic phase transition of graphite into diamond. This process occurs in the diamond window (Fig. 4). The low pressure modification on the other hand is formed by polycondensation processes caused by the formation of serpentine, which controls the C:O:H ratios in the fluid thus leading to the formation of micro diamonds by the presence of chemical reactions.

49

Chapter 7: Volatiles around Totally Embedded Micro Cracks

7.1 Introduction

Earth is considered to be a water planet. The surface processes of the planet are dominated by oceans of liquid water. The water on the planet’s surface controls weathering processes as well as sediment transport and deposition. In the interior, water fluxes melting and controls the solid-state viscosity of the convecting mantle and thus controls volcanism and tectonics. Oceans cover 70% of the earth’s exterior but it only makes up about 0.025% of the planet’s mass. Hydrogen is the most abundant element in the cosmos but it is one of the most poorly constrained chemical compositional variables in the bulk earth (Keppler and Smyth, 2006).

The earth’s crust and mantle partly consist of nominally anhydrous minerals and a large amount of these minerals can incorporate assessable quantities of hydrogen. Nominally anhydrous minerals are minerals that contain oxygen as the major anion, the principal incorporation mechanism is hydroxyl, OH-, and the chemical component is equivalent to water, H2O. The hydrogen proton is a monovalent cation, even though it does not occupy the

same structural position in a mineral structure as a normal cation. The hydrogen proton rather forms a bond with the oxygens on the edge of the coordination polyhedron. The amount of hydrogen incorporated into the structure is sensitive to pressure changes and thus increases with pressure and sometimes temperature. The solubility of hydrogen in nominally anhydrous minerals is thus much more sensitive to pressure and temperature changes than that of other elements (Keppler and Smyth, 2006).

The amount of hydrogen incorporated into the nominally anhydrous phases in the mantle rocks may constitute the largest water reservoir in the planet. This is due to the fact that the mass of rocks in the interior is much larger than the mass of the oceans. Understanding the behaviour and chemistry of hydrogen in minerals at the atomic scale is thus critical in understanding the geology of the planet. Advances in measurement, detection and location of hydrogen in the nominally anhydrous oxide and silicate minerals that compose the planet, have been made in recent years. There have also been advances in experimental methods for measurement of hydrogen diffusion and the effects of hydrogen on the phase boundaries and physical properties whereby the presence of hydrogen in the interior may be inferred from seismic or other geophysical studies (Keppler and Smyth, 2006).

50 To further understand the importance, location and diffusion profiles of hydrogen in the mantle, eclogites from the Roberts Victor mine have been studied. Eclogites can act as water and CO2 reservoirs (Potgieter, 2009). The water and CO2 is found on one-dimensional defects

such as point defects and two-dimensional defects such as cracks and monomineralic and interface grain boundaries (Sommer et al., 2008). These defects are found in garnets and omphacites, which are the rock-forming minerals in the studied eclogite. Water has a major influence on the ductile-brittle regime in the Earth’s mantle (Regenauer-Lieb et al., 2001). A problem that exists in quantifying water and CO2 concentrations is the necessary high spatial

resolution in nanoscale, which is needed for doing measurements of volatiles on point defects. Therefore the exact mechanism of water and CO2 located on point defects cannot be observed

directly. Heggie (1992) described the pumping of water along two-dimensional defects such as cracks and grain-boundaries which are enhanced through deformation. The basics of this process were described by Griggs (1967) and is called hydrolytic weakening.

Hiraga (2004) has found, that the water concentration in the direction of grain-boundaries, are characterized by the increase of incompatible elements towards the grain-boundary. The increase of water in the direction of monomineralic and interphase grain boundaries was described recently (Sommer et al., 2008). These authors used Synchrotron based FT-IR, because normal bench-top FT-IR equipment has a too low sensitivity and spatial resolution to detect water towards grain boundaries. A second problem in quantifying water and CO2

concentration in minerals is that during the eruption of the eclogite from the earth’s mantle to the surface, water and CO2 can be added or removed, depending of the solubility of volatiles

in the kimberlitic melt.

These problems have been solved by measuring volatiles in totally embedded monomineralic grain boundaries in garnets of the studied eclogites. Conventional bench top FT-IR microscopes have an unfavourable trade-off between the brilliance of the IR source and the analysed size of the measured area. This problem has been overcome by using synchrotron based FT-IR. The ANKA synchrotron at the Forschungszentrum Karlsruhe, Germany, with a defraction-limited IR edge-radiation from its 2.5 GeV, 200mA beam has been used for the measurements.

51

7.2 Analytical Methods

The studied samples E18, E19 and E24 are double polished thick sections from the eclogite xenolith (Fig. 6). Totally embedded micro cracks in garnet and serpentinized monomineralic and interphase grain boundaries were studied by using a petrographic microscope as discussed in chapter 3.

The micro cracks in the garnet were studied by using thick sections, which have been polished on both surfaces in paraffin to avoid any contamination with molecular water. The crystallographic orientation was neglected due to the cubic nature of the garnet. The thickness of the analysed double polished sections is 65 μm. The Beer-Lambert law was used to calculate the water content from the FT-IR spectra. An absorption coefficient 0.7194 was used to determine the water content in the garnet grains (Bell and Rossman., 1992). The OH peak was normalized to 1 cm. The study focused on totally embedded micro cracks and serpentinized grain boundaries. Totally embedded micro cracks from the sub-grain boundaries were selected. These fully embedded cracks were selected to avoid any interference from surface cracks or intersections with the polished surfaces of the garnet grain. The fully embedded cracks were investigated with FT-IR in transmitted-light mode. IR absorption spectra in the range from 600 to 10000 cm-1 were acquired at the infrared beam line of the ANKA synchrotron with incident light polarized using a Bruker IFS 66v/S spectrometer coupled to an IRscopeII microscope with a 36x, 0.5 N.A. Schwarzschild objective and a liquid-N2 cooled MCT detector. The spectral region of interest for this study

is the OH stretching range at around 3780 cm-1_{. Brilliance advantage (photon flux per unit}

source area into unit emission angle) of synchrotron light compared to conventional sources was exploited, which allows much higher measurement beam intensity through small sample areas. In addition to the single-spot measurements the synchrotron infrared light of ANKA was used to analyse large numbers of overlapping spots using a step size of 2 μm and an aperture of 4 μm diameter in a grid pattern accessed by an automated X-Y stage.

7.3 O

7.3.1 C

Figure 4 the stud micro c garnets, volatile measure The ana with inc light of and an The siz dry area opening surface inside th 42e, 42

H and C

Correlation

42. Two dim died garnets crack as we , D: CO2 d es present a ements with alysed garn clusions of f ANKA wa aperture of ze of the ana a in the gar g of about ~ (Figs. 42a he crack. Th f). These w

O

2

Distri

n of OH a

mensional m s. A: The to ell as micro diffusion pro around the h a stepwise

net from sam micro diam as used to an f 4 μm in d

alysed map rnet. The inv

~ 5-10 nm & 42c). In he zone aro wet clouds a

ibution A

and CO

2

D

maps showi op surface o diamonds ofile, E: OH crack. All e of 2 μm an mple E18 s monds in th nalyse 512 diameter in is 40 x 40 vestigated m (Figs. 42b) figure 42b ound the cra are about 20

Along To

Distribution

ing the distr of the garne from samp H diffusion maps are 4 nd 512 scan shows total e studied cr scans of ov a grid patte μm (Fig. 4 micro crack ). This crac small inclus ack forms sp 0 microns in

otally Em

n Profiles

ribution of O et sample, B le E18, C: n profile and 40x40 μm ns (Potgieter lly embedde rack (Fig. 4 verlapping s ern accessed 42b). The ba k is approxim k does not sions of mic pots of C:O n diameter a

mbedded M

OH, CO2 an B: Middle su Bottom sur d F: diffusi in size and r, 2009). ed micro cr 42b). The sy spots using d by an aut ackground mately 40 μ crop out at cro-diamon :H bearing and the pea

Micro Cr

nd other vol urface cont rface of the ion profile d consist o racks in the ynchrotron a step size tomated X-Y was measur μm long and t the top or nds can be o volatiles (F ak at ~ 3700 52

racks

latiles in taining a e studied of other f 21x21 e garnet infrared of 2 μm Y stage. red on a d has an r bottom observed Figs. 42d, 0 cm-1 _is

53 indicative of the peak shift of OH in the garnet located in the crystal lattice towards the crack (Fig. 42b).

Two peaks were observed in the studied garnets, one at 3700 cm-1 which is suggestive of the hydro grossular component in the garnet and the second is another at 3655 cm-1 which is typical for the pyrope component in the investigated garnet (Fig. 42b). The peak at 3700 cm-1 becomes more intensive at the totally embedded crack. The OH concentration in the wet clouds is up to 10x higher and is up to 820 ppm (Fig. 42b).

CO2 concentration around the same micro crack was also measured. At ~2350 cm-1 a strong

CO2 band can be observed (Figs. 42a and 43a) The CO2 distribution around the crack shows

nearly the same pattern as the OH distribution and shows similar diffusion rates around the investigated crack (Figs. 42a & 42b). Strong signals have been found between 1360 cm-1 _and

1560 cm-1 in the wavelength bonds (Fig. 42c). These wavelengths bonds can be allocated to those of CO, CH4, CH2O and CH3OH (Fig. 42c and 43c). The wavelength at 1423 cm-1 of the

FTIR spectra, and the Raman shift at 1324 cm-1 are indicative for micro-diamonds (Fig. 42c). The distribution pattern of CO, CH4, CH2O and CH3OH is the same as those of OH and CO2

Figure 4 volatile CH3OH 43. A: CO2 p peak at 136 (Potgieter, 2 peak at 2300 60 – 1560 cm 2009) – 2400 cm-1 m-1_{. Exampl} 1_{, B: OH pea} les of C:O:H ak at 3580 – H bearing vo 3740 cm-1 _a olatiles are C nd C: C:O:H CO, CH4, CH 54 H bearing H2O and

7.3.2 N

Figure 4 sample E well but OH diff stepwise The stu (Fig. 44 investig nm (Fig (Figs. 4 44e). Th μm wid OH loca As in sa and a py more in 10x hig

Non-corre

44. Two dim E19. A: The t no micro d fusion profil e of 2 micron udied garnet 4b). The sa gated micro g. 44b). Th 44a, 44b & hese OH “c de (Fig. 44e ated in the c ample E18, yrope peak ntensive at t gher (Fig. 44

lation of O

mensional ma e top surface diamonds, C: e. All maps n and 512 sca t grain in sa ame analytic crack is ap he investigat 44c). “Spot clouds” are e). The locat crystal lattic two peaks at 3655 cm the embedde 4e).

OH and C

aps showing of the garne : Bottom sur s are 40x40 ans (Potgiete ample E19 s cal setup w pproximatel ted sample ts” of volati also found tion of the p ce at the cra could again m-1 were foun ed crack. Th

O

2

Distrib

the distribut et sample, B rface of the μm in size er, 2009). shows totall was used as y 60 μm lon is also a to iles can be to be paral peak at ~ 3 ack. n be identif nd. As in sa he OH conc

bution Pro

tion of OH a : Middle sur studied garn and consist ly embedde in the mea ng and has otally embe observed in llel to the cr 700 cm-1 is fied. A hydr ample E18 t centration in

ofiles

and CO2 in th rface contain nets, D: CO2 of 21x21 m ed, inclusion asurement o an opening edded crack n the micro rack and are

indicative ro grossular the peak at n the wet cl he studied g ning a micro diffusion pr measurement n free, micro of sample E g of between k as in sam crack (Figs e approxim of the peak r peak at 37 3700 cm-1 b louds is aga 55 garnets of crack as rofile, E: ts with a o cracks E18. The n ~ 5-10 mple E18 s. 44d & mately 25 k shift of 700 cm-1 becomes ain up to

Measur CO2 ba does no (Figs. 4 Figure 4 The figu and inter rements wer and can be o ot show any 42d & 42e). 45. OH profi ure shows tha

rface structu re also don observed at y correlation ile measured at OH is loca ures and does

ne on the C t ~2350 cm -n with the d over a dista ated in two-d s not occur in CO2 concent -1_{. The CO} 2 OH distribu ance of 220 dimensional d n the crystal tration arou 2 distributio ution, which μm X 6 μm defect structu lattice.

und the mic on around th h was the c of a micro c ures which a cro crack. A he embedde case in sam crack of sam act as mono m 56 A strong ed crack mple E18 mple E24. mineralic

57 A profile of about 220 μm was measured at the garnet grain of sample E24. This profile was measured at the lower middel section of the studied rock sample (Fig.7). 256 scans of overlapping spots were made to construct the profile over the totally embedded crack in the garnet (Fig. 45a). A step size of 3 μm and an aperture of 4 μm in diameter in a grid pattern accessed by an automated X-Y stage were used (Fig. 45a). The size of the analysed profile is 220 x 10 μm (Fig. 45a). The background was again measured on a dry area in the garnet. This micro-crack is 60 μm long and is about ~10 nm wide (Figs. 45a). The investigated crack does not crop out at the top or bottom surface and is thus totally embedded (Figs. 45a). The OH is located at the 2-dimensional defect structure at a wavelength of ~ 3700 cm-1 (Figs. 45b & 45c). The carbon bearing volatiles such as CO2, CO, CH4, CH2O and CH3OH could not be